Legal Document Classification For A Leading LPO Using AI

About the Client

Legal Process Outsourcing company in the US. One of the top ranked LPOs as per Frost and Sullivan.

Business Requirement

  • Daily monitor 2600+ US Federal Tax and Legislation websites
  • Download the content and Identify changes in the content
  • Manually go through the content and classify as “Relevant” or “Non-Relevant”
  • A team of 12 lawyers performing this task daily

Our Solution

  • Ellicium developed a bot to scan all the sites
  • Python was used to scrape and download the content
  • Machine learning algorithm- Naive Bayes was used to classify the documents as relevant and non-relevant

Process Flow

Business Outcomes

  • Timely and Real time data was captured
  • 6 person weeks of effort saved every week
  • Dependency on the SME reduced by 60%
  • 80% growth in website analysis